Seeing their current trajectory, LTT would be better off cutting their losses, killing the lab and abandoning benchmarks altogether because its only a matter of time before some colossal mistake makes Intel / Nvidia / AMD angry.
I see it as a damned if they do, damned if they don't sort of scenario. LTT is one of the 'Culture' tech reviewers out their as you point out, where the presumption of their expertise carries a lot of the weight that their actual actions and data fail to back up. What's important here is that these people need to retain that atmosphere of "knows what they're talking about". Through regular issues and mistakes the image doesn't get marred much, and people forget. But what kills this sort of stuff is when you make such a colossal fuckup on a major product launch that you end up making a LOT of enthusiasts waste a LOT of money on a subpar product, publicly, while everyone else is drawing attention to it. The narrative becomes less about the fucked up product, and more about your tech incompetence - Think back to that infamously bad computer build everyone memes on years back. Instant kill of any reputation you might have.
The Labs, as I see, is the only real opportunity they have to avoid, or get ahead of this before it happens. Either the Labs provides enough genuine technical knowledge and experience that they avoid a quality shatter moment during a Nvidia XX90 series launch, or the lab at least provides a smokescreen of "We didn't YOLO it, we have serious people this really was just a mistake". Without the Lab, its not a question of if LTT nukes their current presumption of expertise, but when, and they've already come damned close many times, most recently being GN triggering a lot of people to critically analyze their work.
If they ditch the labs, they ditch this last failsafe for their presumption of expertise. If they don't ditch the labs, I agree in part that its likely they'll eventually misstep and piss off a major player - Although I suspect what'll actually happen is they'll try to pull a GN and be "hard truth seeking" on something with the lab, get over aggressive about it and make way too big of a deal out of a normal failure state/mode/etc, and end up losing their access journalism to one of the big players. If you like watching shitty rushed reviews, just wait until they're trying to review an overlapping new hardware release scenario where they couldn't get anything weeks early.